Back to directory

Elastic Agent Skills

Elastic Agent Skills is Elastic's official library of Agent Skills that teach an AI coding assistant how to correctly use Elasticsearch, Kibana, Elastic Observability and Elastic Security, compatible with agentic IDEs such as Cursor, GitHub Copilot, Windsurf and Gemini CLI.

Elastic Agent Skills

Elastic Agent Skills is built by Elastic and follows the open Agent Skills standard: each skill is a self-contained package with a SKILL.md file of metadata and instructions that an agent reads to complete a specific task correctly, rather than relying on the model's general, potentially outdated knowledge of the platform.

The library's current scope covers interacting with Elasticsearch APIs for search, indexing and cluster management, building and managing Kibana content such as alerts and connectors, and patterns for Elastic Observability, Elastic Security and Agent Builder. Example skills include onboarding an Elastic Cloud organization, configuring its CLI and API keys, and provisioning or operating Elastic Cloud infrastructure such as serverless projects, network security traffic filters, and the lifecycle of hosted deployments.

It is written in JavaScript and marked as a technical preview, with skills expected to change as they gain more rigorous evaluation.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.